Supply restoration in distribution networks using a genetic algorithm

The distribution network supply restoration problem is to reconfigure the network, after a line removal due to fault, such that loss of supply to customers is minimised, while ensuring that current loading and voltage excursions do not exceed permitted levels, and minimising the network loss at minimum switch operation cost. A genetic algorithm is employed to solve this multi-objective constrained optimisation problem. Instead of the standard binary string, the 'integer permutation' encoding scheme is adopted with each integer representing one controllable switch. Graph theory is employed to decide the final status of each switch according to the radiality constraint. The objective function of the GA incorporates all the objectives and constraints required for the supply restoration scheme. A special gene is introduced to search for optimal load shedding in the situation where it is necessary. The proposed algorithm is tested on a 114-switch system that is part of a practical distribution system, and promising results have been obtained. (C) 2002 Elsevier Science Ltd. All rights reserved.